Modern day business machines include data terminals and Electronic Cash Registers (ECR) in which data in the form of binary bits is stored in memory units such as Random Access Memory (RAM) members and which utilize a central processing unit (CPU) for controlling the various functions of the machine. The CPU can be constructed with a computing capacity as to be capable of performing diagnostic routines on the ECR whenever a breakdown of the machine occurs during its operation. But in order to provide this capacity, the cost of the CPU is high thus limiting their use in only the more expensive ECR's. In those ECR's which are designed to be of low cost construction, the CPU is selected by necessity to be of low cost but still capable of controlling the various operations of the ECR for which it was designed. This has resulted in the use of a CPU which does not have the capacity to store and carry out the test routines of the more expensive CPU. To overcome this problem, plugable Read Only Memory (ROM) members have been used to provide the additional diagnostic routines needed for the testing of the ECR. While this method solved the diagnostic portion of test programs, the data storage area in the CPU could not be tested because of the limited storage capacity of the low cost CPU. In addition, it was found to be desirable if the operator could observe the results of each test routine so as to eliminate further testing of the machine once the defect has been discovered.